November 7, 1991 (revised August 28, 1992)

In the United States, the most common dental restorative material is amalgam containing about 50 percent mercury at the time of mixing with various silver alloys. It has been used for more than 150 years. However, because of improved dental hygiene, this is the first generation of people who have reached the upper ages and who have also preserved their teeth (1), thus increasing the length of exposure to dental restorative material. From 1971 to 1974, U.S. adults 18-74 years of age had an estimated average of 6.9 filled teeth (2).

Exposure to mercury from amalgam occurs through several avenues: inhalation of air containing elemental mercury released from the amalgam; ingestion of amalgam particles abraded from restored surfaces and of saliva into which both elemental and corrosion produced inorganic mercury products have dissolved; and ingestion of amalgam particles generated during dental restorative procedures (placement, restoration, or removal) (3-5). Another source of exposure to elemental and inorganic mercury is "tatooing," which may be created when, as amalgam is being removed, amalgam particles are physically embedded in soft tissue adjacent to the restoration area (6).

Mercury in its many forms is widely distributed in the environment and trace levels are present in air, water, and food.

The toxicity of mercury and its compounds, recognized since antiquity and widely acknowledged in industry, has recently been reviewed (7-12). Signs and symptoms associated with mercury intoxication from elemental mercury include tremor, ataxia, personality change, loss of memory, insomnia, fatigue, depression, headaches, irritability, slowed nerve conduction, weight loss, appetite loss, psychological distress, and gingivitis (7,913). Most of these signs and symptoms have been associated with persons with long-term occupational exposure to air concentrations of mercury greater than 50 µg/m³ whose urinary mercury concentrations are greater than 100 ug/L. Clinically significant effects (erethism, intention tremor, gingivitis) have not been reported below air concentrations of 100 µ(g Hg/m . Most effects observed in persons exposed to mercury in air concentrations below 100 ug Hg/m³ are preclinical e.g., slowed nerve conduction, short term memory loss, special instrumental tests for tremor. No clinical findings on kidney function decrement have been found in persons exposed to air mercury concentrations below 100 mg Hg/m³ . In comparison the range of mercury in urine for persons with no clearly identifiable occupational source of mercury exposure is up to 20 ug/L.

Because of the known toxicity of mercury, various agencies have developed limits for mercury vapor in workplace air to protect the health of workers. These limits are based on the assumption that the workers will have only 40 hours per week exposure to these levels of mercury vapor. Both the Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) (14) and the National Institute for Occupational Safety and Health (NIOSH) recommended exposure limit (REL) (15) for mercury vapor in the workplace are 50 micrograms per cubic meter (µg/m3) as a time-weighted average (TWA).

NIOSH in its criteria document (16), noted three areas that make selecting a specific workplace level for mercury difficult. These were 1) the prevalence in the general population of the signs and symptoms identical to those associated with early signs of mercury intoxication, 2) sampling and analysis difficulties that complicate the validity of measuring air levels, and 3) a lack of established methods to specifically identify the effects of exposure to low levels of mercury. On the basis of available data and their recognized limitations, NIOSH stated that a specific level at which a standard should be established cannot be identified; but concluded that the standard should not exceed 50 ug/m³.

The World Health Organization (WHO) adopted a health-based recommended lim3it for occupational exposure of 25 ug/m³ . The WHO Study Group selected this value to ensure a reasonable degree of protection not only against tremor but against mercury-induced nonspecific symptoms (17). Effects induced by exposures that exceed these levels have been well documented (7,9,10,16). Long-term exposure to elemental mercury in workplace air has produced tremor in some workers at mercury levels of 100 ug/m³ and nonspecific symptoms at levels of 50 ug/m³. In the United States, there is little difference between acceptable workplace exposures and those that produce symptoms after long-term exposure. Therefore, basing low-level or ambient exposure on some fraction of the permissible workplace exposure is not likely to provide satisfactory protection for the sensitive portion of the population such as the young, the aged and the chronically ill.

The Agency for Toxic Substances and Disease Registry (ATSDR), through its Toxicological Profile for Mercury (11), developed 0.3 ug/m³ as a chronic inhalation minimal risk level (MRL) for long-term human exposure to mercury vapor in ambient air. The basis of this value was a study of workers (18) exposed to a time weighted average of 26 ug/m³ for an average of 15.3 years with an increase in intentional tremor compared to the control group. Using this exposure value as a lowest observed adverse effect level (LOAEL) and a safety factor of 100 (10 for using a LOAEL and 10 for human variability), ATSDR derived the MRL. The MRL is defined as the level of mercury vapor below which a person can be continuously exposed with no harmful health effects.

The Environmental Protection Agency (EPA) is developing an inhalation reference concentration (RfC)(l9) for elemental mercury, the summary of which will be in EPA's Integrated Risk Information System (IRIS) for elemental mercury by late summer 1992. The RfC represents a health-based risk estimate of a daily inhalation exposure to human population (including sensitive subpopulations) that is likely to be without deleterious non-cancer effects over a lifetime exposure. The value of the RfC (0.3 ngm) is the same as the MRL developed by ATSDR.

In this document, we evaluate exposure to mercury from dental amalgam restorations through a review of the significant literature that describes the evidence for possible health effects produced from exposure to mercury from this source.

Mercury Forms, Intake, Uptake, Metabolism, and Excretion

Knowledge of the uptake routes and distribution of mercury is essential to the critical interpretation of data on levels of mercury in body tissues and fluids. Mercury occurs in three chemical forms: elemental (valance 0); inorganic (valance +1 and +2); and organic, including alkyl (e.g., methyl) and aryl (e.g., phenyl) mercury compounds. Each has different physical and chemical properties, different rates of absorption and excretion, and different distribution patterns. These differences effect the toxicological effects of the various forms of mercury.

In recent articles investigators have reviewed human intake and retention of environmental mercury(10,12). Vapor is the predominant chemical form of elemental mercury found in air (20,21). While elemental mercury has a high vapor pressure, which can result in a concentration of about 50 mg Hg/m³ at 37° Celsius, the concentration in ambient air is quite low.

Methylmercury is produced by anaerobic bacteria interacting with inorganic mercury in marine and freshwater sediments. Monomethylmercury is believed to bind to edible proteins in fish muscle. Because its excretion by fish is extremely slow, the aquatic food chain plays a major role in the accumulation process. Such accumulation contributes to the mercury body burden of people who frequently consume seafood products (8,10-12). Methylmercury is rapidly absorbed into the blood from the intestine, with about 90 percent of that present in the blood found in the red blood cells (25).

Dietary methylmercury does not appreciably elevate urine mercury levels, because about 90 percent of it is excreted in the feces (10) with excretion beginning as a biliary secretion. The methylmercury in bile is almost completely reabsorbed in the intestine, forming an enterohepatic cycle. Studies have shown (26,27) that it is not until intestinal flora converts methylmercury to Hg++ that mercury from this form can be excreted in the feces. This is one reason why urine mercury is useful in evaluating occupational exposure to inorganic mercury (elemental vapor and inorganic salts).

In many studies of mercury exposure from amalgam dental restorations, investigators report results as total blood mercury. These data blur the distinction between elevations of the inorganic mercury in blood (due to exposure to mercury vapor from amalgam restorations) and organic mercury in blood (due to dietary exposures, usually methylmercury).

When high exposure has occurred (e.g., industrial, seafood consumption), the elevation of total blood mercury accurately reflects exposure, simply because the predominant quantity of mercury in the blood, be it organic or inorganic, is due to that high exposure. Unfortunately, this is not the case in low-level exposures, such as to mercury released from amalgam dental restorations. Although some investigators have reported that bacteria obtained from the oral cavity have produced methylmercury in vitro (28), Eley and Cox (29) found no evidence of its having been produced in situ.

In this evaluation, we do not address toxicity or the general metabolism of nonelemental forms of mercury (Hg, Hg and organic) except when such information may aid in an understanding of the metabolism, mechanism of action, storage, or excretion of elemental mercury.

Elemental mercury, the focus of this document, makes up about 50% by weight of dental amalgam when it is mixed. Atoms of elemental mercury continuously diffuse from the amalgam used as a dental restorative (30-33). This molecular diffusion proceeds from the amalgam through the amalgam oxide, the saliva, and the air boundary layers into air flowing through the mouth. Each of these layers provides some resistance to the movement of elemental mercury from the amalgam into the mouth air flow. Chewing, brushing, and other abrasive stress on exposed surfaces of dental amalgam can alter the protective characteristics of the oxide layer and increase the rate at which elemental mercury is released (30-32).

In addition to the release of elemental mercury to the mouth air, amalgam also releases inorganic mercury (Hg⁺⁺) as a corrosion product (29,30). This has also been demonstrated in vitro in both natural and synthetic saliva (30). Both elemental and inorganic mercury are introduced into the gastrointestinal tract in saliva swallowed by persons with mercury containing restorations. Under the aerated conditions of the upper GI tract however, almost all the elemental mercury will be converted to inorganic mercury (Hg ) (34).

During his 24-hour study, Berglund (33) collected unstimulated saliva for about 40 minutes from each of his subjects. He measured release rates of total mercury into saliva that ranged from 1.8 to 13.8 ng Hg/min. Making the assumption that this mercury release rate from amalgam to saliva is constant throughout the day, these data give a range for mercury ingestion from this source from 2.6 to 19.9 µg for 24-hours. Gastrointestinal absorption of inorganic mercury (Hg⁺⁺) is no more than 10 percent (12) and the absorption of elemental mercury by this route is also very low (9). Intestinal absorption of inorganic mercury is much greater than that for elemental mercury. Thus, assuming all the mercury represented by the Berglund data was inorganic mercury would yield the greatest absorption and would give an uptake range of about 0.18 to 1.4 µg/day.

Mercury vapor is absorbed rapidly into the bloodstream and distributed to all major organs and tissues (20); mercury is most highly concentrated in the kidney, with other organs (brain, lung, liver, gastrointestinal tract, and exocrine glands) also showing varying degrees of elevated concentration (9,10). Elemental mercury in blood can cross the placenta and the blood-brain barrier, it can then be oxidized to inorganic mercury (Hg+ ), which has a limited ability to cross biological membranes. Thus, mercury can be retained in brain (35) and fetal tissues (36,37).

Recently, mercury metabolism in the body has been reviewed (9), but it is still not completely understood. Because initially, mercury accumulates in the kidney as a complex with metallothionein, a sulfhydryl rich compound produced in the kidney, a delayed increase in urinary excretion occurs after exposure.

Cherian and coworkers (38) studied the excretion of mercury in five human volunteers after a single exposure to vapor at 100 ug/m³ for 14 to 24 minutes. Using ¹⁹⁷Hg and ²⁰³Hg to trace the distribution and release of mercury from the body, the investigators found that during the first week after exposure about 11 percent of the absorbed dose of mercury was excreted, predominantly in the feces. During a skin absorption study, Hursh and coworkers (39) found that this predominance of feces-to-urine excretion continued for 10 to 20 days after a single exposure to mercury vapor. Tejning and Ohman (40) found in workers exposed to mercury vapor for long periods that urinary excretion exceeded fecal excretion by a 60-to-40 ratio.

Hursh and coworkers (20) measured the clearance of mercury vapor in five human volunteers, using radioactive mercury. These investigators found a biphasic clearance, with the slower fraction having an average whole-body half-life of 58 days for these subjects. Because of the physical half-life of the isotope used and the low dose, these investigators could not evaluate the possibility of an even longer half-life phase in the clearance of mercury from the body.

There is also evidence that mercury vapor is excreted in the exhaled breath after inhalation exposure. Hursh coworkers (20) found that about 7 percent of the absorbed dose of inhaled mercury vapor was exhaled within three days. On the basis of the half-life of mercury exhalation, these investigators believed that loss of mercury by exhalation would be insignificant within 1 week of a single exposure (38). Because persons with dental amalgam restorations continuously inhale mercury vapor, these data suggests that their exhaled breath always contains a small quantity of mercury vapor excreted from their lungs. However, this contributes only a very small amount to the total mercury vapor present in the exhaled breath of persons with dental amalgam restorations.

Elemental mercury dissolves in lipids and readily diffuses across cell membranes. Once within the cell, it is oxidized to inorganic mercury (Hg⁺⁺ ) by catalase enzymes (41) present in red blood cells, brain, liver, lung, and probably the cells of all other tissues. Inorganic mercury (Hg++) readily binds to tissue ligands, which may explain why it accumulates in various tissues. Kosta et al (42) found a nearly 1 to 1 molar ratio between mercury and selenium for those organs which accumulate and retain mercury. This relationship may represent a chemical form that effectively removes mercury from the normal biological turnover and hence play a significant role in mercury accumulation for those organs.

Eventually the distribution of mercury in body tissues obtained by inhaling mercury vapor approaches the distribution of mercury obtained by ingesting inorganic mercury (Hg +), except that levels in the brain are greater when the mercury is inhaled. Animal data (43,44) show that upon initial exposure the amount of mercury in the brain is about 10 times greater when the exposure is to elemental mercury vapor than when exposure is to inorganic mercury (Hg+ ). Human autopsy data (45) show that mercury concentrations in brain tissues are greater in persons with amalgams compared with concentrations for those without them. The human autopsy data, however, represent the accumulation of mercury from long-term, low-level exposure to all forms of mercury.

Evidence suggests that the neurological effects produced by exposure to mercury vapor results from its oxidation to the divalent mercury ion in brain tissue and the ensuing interaction of these mercuric ions with enzyme sulfhydryl groups thereby inhibiting their function (46).

An example of the role played by enzyme inhibition in mercury accumulation is ethanol's ability to inhibit catalase oxidation of elemental mercury (47,48), which results in decreased blood mercury concentrations (49,50) as well as lower tissue levels (51) in persons inhaling mercury vapor. Ethanol also enhances reduction of Hg⁺⁺ to elemental mercury both in vitro (52,53) and in viva (54,55). Clarkson and coworkers (10) postulates that ethanol inhibits catalase activity, which oxidizes elemental mercury, thereby shifting the equilibrium toward elemental mercury and increasing the amount of mercury available to migrate through the cell membrane. This postulate has been supported by in vitro data which show that other catalase inhibitors also increase the volatilization of mercury vapor from tissues (52).

Human Exposure to Mercury from Dental Amalgam

Within the past decade, investigators have demonstrated higher levels of mercury vapor in the oral cavities of persons with amalgam restorations compared to with levels for persons without such fillings. Mercury vapor in humans has been sampled in exhaled breath (32,56,57), in the oral cavity with the mouth open (58 60) or closed (61,62), and through a catheter placed in the trachea via a bronchoscope (62). These data suggest (59) that mercury is continuously released in the oral cavity from amalgam dental restorations. The rate release is dependent upon many factors including: area, age, and composition of the amalgam, as well as the quality of the surface oxide layer.

Numerous investigators have demonstrated increased intraoral mercury vapor concentration especially after occlusal surfaces were stressed by chewing or tooth brushing (32,33,56,57,59,61,62). Intraoral mercury vapor levels were directly correlated with the number of amalgam fillings(56,57,61). A positive correlation was obtained between surface area or number of amalgams and mercury levels in body fluids (59,61-64) and human autopsy tissue samples (51,45). Estimates of the daily uptake from the inhalation of mercury vapor from dental amalgams suggest that amalgams may be the largest source of mercury in persons without occupational exposure (10,65,66). Details relevant to the routes, rates of uptake, distribution patterns of mercury, and the long-term health effects from or presence of dental amalgams are incomplete.

Factors Affecting Estimates of Daily Intake of Mercury Vapor From Dental Amalgam

Estimating daily intake of mercury from dental amalgams requires a knowledge of both the rate and duration of vapor production. Most data on intraoral levels of mercury vapor are based on a small series of samples after controlled stimulation of the amalgam surface. Several studies focused on mercury levels obtained after stimulation by gum chewing or tooth brushing as an estimate of the stimulated rate for three meals and three snacks per day. Data by Berglund (33) suggest, however, that such estimates are high because the rate of mercury vapor production following gum chewing exceeded that following a typical meal. Bergland (59) also found that the amount of mercury produced from amalgam restorations into the oral cavity is essentially independent of air flow over a wide range of flow values.

Fifty percent of the intraoral mercury vapor is the upper limit of the range used to estimate the amount inhaled (the rest is lost to exhalation). The lungs absorb about 80 percent of mercury vapor present in inhaled air (20,21). Most of this mercury vapor diffuses directly and rapidly across the alveolar membrane into the blood, and only a small fraction is deposited in pulmonary tissues.

Any corrections for breathing style should include the fraction of time spent inhaling through the mouth. In the studies reviewed, investigators estimating exposure have usually included corrections for the ratio of time spent in oral respiration and the percentage of lung absorption. However, the values used by some of these researchers for the ratio of oral to nasal breathing may not be appropriate. Niirumma et al. (67) have shown that a split in air flow occurs in normal augmenters (people who normally breath through their nose but augment ventilation by employing mouth breathing at high ventilation rates) when the minute ventilation rate exceeds approximately 35L/min At 35 L/min, the nasal portion of the minute volume decreases to 57 percent of the total minute ventilation.

Intraoral Mercury Vapor Production and Estimation of Daily Intake

Several studies have measured the mercury vapor released from amalgam dental restorations before and after gum chewing or brushing. On the basis of these measurements, the investigators and others have estimated the daily absorption of mercury vapor by inhalation. Table 2 (next page) presents these estimates of daily mercury absorption from dental amalgam restorations developed by the researchers whose work has been reviewed in this document. These estimates range from 1.24 to 27 µg/day.

Svare and coworkers (56) measured mercury levels in a 3 liter exhalation study of 40 subjects with up to 21 amalgams and in 8 subjects without amalgams. The investigators found that resting levels of mercury vapor were higher in the oral cavities of those with amalgam fillings (means 0.88 ug/m³ versus 0.26 ug/m³). They observed a fifteen-fold increase (mean 13.74 µg/m3) in oral mercury levels in those with amalgams after 10 minutes of stimulation by gum chewing. The mercury vapor levels in expired air for those subjects without amalgam fillings decreased to 0.13 ug/m³ after a similar period of chewing. For patients with amalgam fillings, the increase reported was proportional to the number of their amalgams.

Abraham and colleagues (61) assessed intraoral mercury levels in 47 subjects with and in 14 subjects without dental amalgams. The investigators used a mercury detector to measure mercury in air from an external source flushed through a two-holed stopper held between the lips while the subject breathed through the nose. Before stimulation by chewing gum, those with amalgam fillings had a mean intraoral mercury measurement of 2.24 ng/15 sec compared with 1.13 ng/15 sec for subjects without amalgams. After stimulation for 3 minutes at a rate of 120 chews per minute, the mean mercury level increased to 18.97 ng/15 sec (an over eight-fold increase) for those with amalgam fillings while the mean mercury level for subjects without amalgam remained unchanged at 1.06 ng/15 sec. Oral mercury levels correlated with the number and surface area of dental amalgams.

Vimy and Loscheider (60) measured intraoral mercury levels among 35 subjects with amalgams and 11 subjects without amalgams. Subjects were asked to refrain from all oral stimuli for 1 hour before the sampling. Sampling consisted of rapidly moving a tube attached to a mercury detector around the open mouth for 20 seconds. The mean intraoral Hg levels for subjects with amalgams was 4.91 ug/m³ before chewing and 29.10 ,ug/m³ after chewing gum for 10 minutes. For subjects without dental amalgams, the unstimulated rate was 0.54 ug/m³, which did not change significantly after stimulation. The total number of amalgam surfaces showed a positive correlation with stimulated mercury levels, whereas the total number of occlusal amalgam surfaces showed a positive correlation with both unstimulated mercury levels and stimulated mercury levels.

In a follow-up study, Vimy and Lorscheider (58) measured intraoral mercury levels at 5 minute intervals during 30 minutes of continuous gum chewing and for 90 minutes thereafter. These data were integrated to obtain the area under the curve. The investigators estimated daily exposure by assuming a respiratory volume of 6 liters per minute, 80 percent absorption of mercury vapor by the lung, an oral to nasal breathing ratio of 50 percent while chewing and 35 percent after chewing, and stimulation periods corresponding to three meals and three snacks per day. On the basis of collected data, mercury generation ascribed to these periods of eating comprised 135 minutes of chewing and 540 minutes of decreasing vapor release following stimulation.

Further, the investigators obtained an estimate of the average intake from amalgams of 19.8 ug/day for all subjects (1-16 amalgams) and 29.24 payday in subjects with 12 or more dental amalgams. Several investigators have documented that Vimy and Lorscheider's assumption relating sample collection rate to physiologic ventilation rate was erroneous (59,62,68). Vimy and Lorscheider's estimate of daily intake of mercury from dental amalgam may be high by a factor of 16 (68-70). Mackert (68) calculated a value of 1.24 µg/day using the data of Vimy and Lorscheider for all subjects and 1.83 ug/day for those subjects with 12 or more dental amalgams.

Patterson and coworkers (57) measured elemental mercury levels in expired air. Subjects exhaled at about 2 liters per minute before and after tooth brushing. Samples were analyzed with a photoaccoustic mercury detector. Of a total sample of 172 subjects (including 5 without amalgams) studied, only 106 (including 2 without amalgams) provided expired air both before and after stimulation. In a subgroup of 104 subjects (2 without dental amalgams) the investigators reported a mean Hg concentration of 1.9 ng/L before stimulation. After tooth brushing for 1 minute, the mean Hg concentration was 8.2 ng/l. In a subgroup of 94 subjects for whom the number of amalgam tooth surfaces was known (but not reported), the correlation coefficient for the number of tooth surfaces versus the concentration of Hg in expired air increased from r = 0.41 before stimulation to r = 0.63 after stimulation. On the basis of the value obtained for the mercury concentration before stimulation for the upper tenth percentile of 172 subjects, the investigators estimated uptake of mercury from amalgams at 27 ug/day. Because of the way the mercury vapor measurements were performed, however, the estimates of Patterson et al. are probably at least a factor of six too high (68).

Using data from several of the previously described studies, Clarkson and colleagues (65) estimated the total daily mercury vapor absorption from dental amalgam restorations by assuming an equivalent of 4 hours per day at a constant, fully stimulated rate, 50 percent loss to exhalation, 80 percent uptake by the lung, 100 percent oral breathing during stimulation and 50 percent oral breathing after stimulation. Using these criteria and data from four published studies they found 2.5 ug (57), 2.9 ug (58), 8.0 ug (61), and 17.5 ug (56), respectively. However, some of the daily dose estimates required major assumptions because the original authors had not supplied crucial information on their experimental methodology, such as the flow rate of exhaled air through the mouth The actual daily doses are probably lower than these estimates.

Langworth and coworkers (62) used an elemental mercury detector to measure intraoral and intratracheal mercury levels after tooth brushing among 10 subjects with a mean of 25 amalgam surfaces. Subjects were instructed to breathe slowIy, but not at a defined ventilation rate. Tracheal concentrations during inhalation were below the instrumental detection limit of 1 ug/m³ for five subjects and ranged from 1-6 ug/m³ for the remaining five subjects. The mean oral mercury level after stimulation was 56.4 ug/m³. To estimate the daily mercury intake from dental amalgams, the investigators assumed a total stimulated period of 4 hours per day a a tracheal air concentration of 2 ,µg Hg/m³, a nonstimulated period of 20 hours per day at a tracheal air concentration of 0.4 µg Hg/m³, a ventilation of 10 m /day, 50 percent oral respiration, and 80 percent absorption of mercury by the lung. These values gave an estimated intake of 3 ug/day.

Berglund (33) measured intraoral mercury among 15 subjects with more than 9 occlusal amalgam surfaces and 5 subjects without amalgams for a 24-hour period during which they followed a prescribed schedule of diet and tooth brushing. These subjects had no prior occupational exposure, and those with amalgams placed within the past year were excluded. Subjects estimated their fish consumption to assess variations in environmental exposure to mercury. Intraoral air was sampled for 2-minute periods every 30 to 45 minutes for 24 hours. Samples were analyzed by atomic absorption spectrometry, and the mercury produced per of unit time was calculated. The area under the time-concentration curve was integrated to obtain the amount released daily for each individual. The

daily intake of mercury from dental amalgams was estimated by assuming 80 percent absorption by the lung and 50 percent loss to exhalation. Oral respiration ratios of 0.4 percent at rest, 58 percent during conversation, and 17 percent during sleep were each applied evenly to one third of the total daily rate of release. Berglund calculated the mean daily dose of inhaled mercury vapor from dental amalgams to be 1.7 µg with a range from 0.4 µg to 4.4 µg. The daily dose was not significantly related to the occlusal surfaces or to the total number or area of amalgams surfaces.

Mercury Levels in Body Fluid

Abraham and coworkers (61) measured blood mercury levels in 47 medical students with amalgam fillings and 14 students without dental amalgams. Morning baseline blood mercury levels were 0.7 µg/L for those with amalgam fillings compared with 0.3 µg/L for subjects without amalgams. Blood mercury levels correlated with both the amalgam surface area and the number of amalgams. Subjects were surveyed for 22 risk factors that might correlate with blood mercury levels, including occupational and home exposure, seafood consumption, chewing habits, smoking, alcohol consumption, and medications. A univariate model was constructed for each of the dependent variables, and the only one of these 22 factors found to correlate with blood levels was teeth grinding.

Snapp and colleagues (64) studied 5 men and 5 women for 4 to 18 weeks (median, 6.6 weeks) to establish baseline blood mercury levels before all dental amalgams were removed. The subjects selected had no current occupational exposure to mercury and reported eating little or no fish or seafood. All the subjects were asked to abstain from eating seafood while enrolled in the study, and all but one complied. The mean baseline blood mercury level across all subjects was 2.18 µg/L before the amalgam fillings were removed. Mercury levels were found to correlate with the number of occlusal amalgam surfaces and the total number of amalgam surfaces. After the amalgam fillings had been removed, weekly blood sampling continued until data analysis demonstrated that each individual's blood mercury concentration had changed from the baseline value with a 95 percent confidence level. The average decrease in blood mercury level, based on the final mercury level from each subject, was 1.13 ug/L. The investigators, using a steady state equation, estimated a daily dose of 1.3 µg mercury from amalgams before removal. Because the post-removal mercury data show a continuing decline, this dose is probably an underestimate of the daily mercury intake.

Molin and coworkers (71) demonstrated the effect of amalgam restorations on mercury levels in body fluids after the restorations had been removed. For subjects studied before and after the amalgam was removed, the investigators determined that removal of this mercury source resulted in a decrease in mercury levels in both plasma and urine. Average plasma levels decreased from 0.9 ug/L, as measured 3 to 4 months before removal to less than 0.5 µg/L 12 months after removal. These researchers also found that the urinary mercury levy decreased during this period from a level of 1.00 µmol Hg/mol Cr measured 3 months before amalgam removal to 0.27 µmol Hg/mol Cr 12 months after removal.

In the study by Berglund (33), a morning urine sample was obtained from 15 subjects with and 5 subjects without dental amalgams. Those subjects with mercury amalgam restorations had an average creatinine-corrected urine mercury concentration of 1.27 umol Hg/mol Cr compared with 0.22 µmol Hg/mol Cr for those without amalgam fillings. In this study, urine mercury levels were not found to correlate with the estimated daily dose nor with the total or occlusal amalgam surface area. Because in nonoccupationally exposed individuals urine mercury concentrations reflect dietary exposure to inorganic mercury (which was estimated but not excluded for this study), further study is needed to assess the contribution of amalgam mercury to urine concentrations and renal load.

Olstad and coworkers (63) studied 59 students with a mean 5.8 amalgam surfaces and 14 students without dental amalgams (5 previously had amalgams that were exfoliated) from the sixth grade class of two elementary schools in Norway. Mercury data were from morning urine spot samples corrected for creatinine clearance. Mercury levels for those with amalgam fillings were higher (0.58 versus 0.17 nmol/mmol Cr) than for subjects without amalgams. Creatinine adjusted mercury levels correlated with quantitative estimates of dental amalgams.

Fosten (72) reported on the measurement of mercury in blood samples from 35 patients at a dental clinic in Finland. These patients had self-reported subjective symptoms suggestive of mercury toxicity, but no clinical diagnosis. Each patient had a minimum of 10 occlusal amalgams. Blood samples, drawn l5 minutes after the subject chewed paraffin for a period of 30 to 60 minutes, were examined by cold vapor atomic absorption spectrometry (AAS) for both total mercury and inorganic mercury levels. The results show an average total blood mercury level for the subjects to be 4.0 ug/L and an average inorganic blood mercury level of 1.79 ug/L. These results showed no increase in the total mercury blood levels compared with an average total mercury blood levels in the Finnish population of 3.7 µg/L. About half the blood mercury in this study group consisted of inorganic mercury, a portion of which probably originated from dental amalgams. Because no efforts were made to limit occupational or dietary exposure, it is difficult to assess what proportion of the exposure might be accounted for by environmental sources.

In a study by Nylander and colleagues (74), samples of tissue from the occipital cortex, pituitary gland, renal cortex, olfactory bulbs, thyroid gland and liver were collected from autopsies of 8 dental staff members and 27 controls. Results of sample analysis from each of the tissues were not reported for each of the subjects. Neutron activation analysis was used to examine samples for total mercury content. Table 5 shows the mercury concentrations from the dental staff and controls (including those without amalgams) for the pituitary gland, occipital cortex, and renal cortex. Autopsied material from the olfactory bulbs showed low mercury concentrations for both groups. A linear correlation between total mercury in the pituitary gland and number of amalgam surfaces was obtained with data from the control group r = 0.53). When the data from individuals in the control group without amalgams and two individuals with employment history of possible mercury exposure were excluded the correlation was improved (r = 0.65, p 0.01).

Nylander and colleagues (51) obtained total mercury levels in the kidney cortex from seven subjects with and five subjects without amalgams. The mean mercury level among subjects with 11 to 33 amalgams was 433 ng Hg/g wet weight compared with 49 ng Hg/g wet weight for subjects without dental amalgams. The difference in mean values was statistically significant by the student's t-test (N = 12, p 0.02). The relation between the number of amalgam surfaces and total mercury in the kidney cortex has a regression coefficient of 0.56 with a two tailed student's t-test of p (N=12). The investigators identified type of food, especially fish, alcohol, and smoking as potential confounding factors. The mean mercury levels included levels obtained from three subjects who were electricians and who therefore may have had considerable occupational exposure to mercury.

Animal Uptake of Mercury From Dental Amalgam

Tests in laboratory animals can provide insight into the possible human response to chemical exposure. However, if the route of exposure, the form of the chemical, or the metabolic processes are different between the animal and man, the results of such tests must be interpreted cautiously.

Hahn and coworkers (76) measured tissue levels of ²⁰³Hg in a 61-kg ewe 29 days after placement of 12 occlusal amalgam fillings containing radioactive mercury. Teeth containing amalgam were removed intact, and a whole-body scintigram was obtained. Tissue samples were weighed at autopsy, and isotope measurements were obtained. The documented tissue levels of mercury are shown in Table 7. The authors identified three uptake routes of amalgam-based mercury as 1) lung and tracheal mucosa, 2) gastrointestinal tract, 3) oral mucosa, tooth root and surrounding bone. The data from this study, however, show only target tissues where mercury localized and not the direction of movement. The principal organs to rapidly accumulate mercury during this time period were the liver and kidney. The stomach also showed substantial mercury accumulation, showing the effect of the large quantity of mercury passing through the gastrointestinal tract, as shown by the mercury concentration in the feces. While biliary excretion of inorganic mercury occurs, the low level or mercury in the blood (9 ng/g) shows that the quantity of mercury available for excretion by this route is small compared to the quantity in the feces (4489 ng/g). Thus, biliary excretion contributes only a small portion to the mercury in the feces of the sheep in this study.

After 29 days exposure to 12 dental amalgam restorations, the sheep kidney contained more than 15 times the mercury found in human kidney samples (51) from persons with dental amalgam restorations (range of amalgam surfaces 11 to 33). However, the sheep occipital cortex contained about the same concentration of mercury (3.5 ng Hg/g) as found in the that tissue from humans (3.8 ng Hg/g) with O to 1 amalgam dental restoration (45). The data from liver, kidney, and feces shows that a large quantity of mercury passed through the blood and gastrointestinal tract of the sheep in only 29 days, but that the brain absorbed only a small fraction.

The mercury accumulation in tissue found in these two studies (77,58) show similar results, with the mercury concentration reaching a plateau in most tissue about 40 days after placement of the amalgam. The sheep kidney showed 10 to 20 times as much mercury compared with concentrations reported for kidneys from humans with many dental amalgam restorations. However, brain tissue mercury concentrations in the sheep were similar to those in persons with few amalgam dental restorations.

Several factors may explain the relatively higher levels of mercury detected in the sheep kidneys as compared to human kidneys. These factors could include 1) the temporal relationship of the study to the placement of 12 occlusal amalgam dental restorations at one time 2) the test animals were followed only briefly 3) the chewing patterns of the remnants (sheep) and the abrasiveness of the remnant diet are very different from those of humans, and 4) the organ distribution of mercury in the sheep appears to reflect gastrointestinal uptake of either elemental or inorganic mercury rather than lung uptake of mercury vapor. Therefore, the results of these studies may not be relevant to humans.

Hazard Identification

A wealth of information (7-13,16) is available on human intoxication by elemental mercury from industrial, inadvertent, or deliberate exposures, both short- and long-term, which have produced clinical signs and symptoms of deleterious health effects. The EPA in its review (12) stated that the scientific literature did not identify a threshold urine value below which mercury has no effect on psychomotor function. However, EPA could not causally link the decrement in psychomotor function to low urine mercury levels.

Acute exposure to high concentrations of mercury vapor can lead to death from pulmonary failure (79-81). However, mortality from acute exposure to mercury vapor is unusual; it occurs primarily in homes, and is associated with the recovery of precious metals. Some investigators have reported neurological effects caused by a brief but intense exposure to mercury vapor (82). In a review, Jaffe and coworkers (83) noted that infants 4 to 30 months of age are more susceptible to mercury intoxication than older children and adults. Among children, exposure to elemental mercury can cause an idiosyncratic reaction known as acrodynia, a syndrome characterized by painful extremities, extreme irritability, and pink rash on the fingers, toes, and nose. An example of this disease was recently reported by Agocs and colleagues (84) in a 4 year-old boy exposed to mercury vapor released by an organic mercury fungicide in the interior latex paint applied to the walls of his home.

In most tissues, elemental mercury is readily oxidized by catalases to inorganic mercury (Hg⁺⁺); therefore, assessing some of the effects of inorganic mercury (Hg⁺⁺) as part of this evaluation is reasonable. The National Toxicology Program (NTP) has performed toxicity and carcinogenicity studies after 2-year gavage studies of mercuric chloride in 344/N rats and B6C3F₁ mice (85). The NTP review of the results of these studies concluded that, on the basis of increased incidence of squamous cell papillomas of the forestomach in male rats, some evidence of carcinogenicity existed. On the basis of the occurrence of two renal tubule adenomas and one renal tubule adenocarcinoma, equivocal evidence existed of carcinogenic activity of mercuric chloride in the male mice. Doses used were 1.25, 5, or 20 mg/kg/day. In the male mice, the 5-mg/kg dose produced mercury concentrations of 36,000 ng/g in the kidney, 3,000 ng/g in the liver and about 100 ng/g in the brain. The kidney concentration reported in these mice was more than 50 times that found by Nylander and colleagues in humans with amalgams. The significance to humans of tumor development in the rat forestomach is not well understood.

In contrast to these findings are those of Cragle and colleagues (86). The authors studied the mortality of a cohort of 2,133 white males exposed to elemental mercury vapor between 1953 and 1963 and "followed" through 1978. The comparison group consisted of 3,260 workers employed at the same plant but without mercury exposure. There was no statistically significant excess mortality in the exposed group compared to the nonexposed group. However, the mortality experience for the comparison group showed statistically significant increased standard mortality ratios (SMR's) computed with the experience of the age-adjusted U.S white male population for cancer of both 1) the lung and 2) brain and other central nervous system CNS tissue. The SMR for lung cancer among the mercury-exposed workers was elevated compared with the U.S. white male population, but the increase was not statistically significant. The investigators concluded that exposure to mercury vapor in the workplace was not related to any excess deaths observed.

Exposure Assessment

Absorption of elemental mercury depends on the route of exposure. Lung absorption of mercury vapor ranges from 65 to 85 percent and depends on both the rate and depth of breathing (20). In the literature we reviewed, all estimates for daily uptake from mercury-containing dental restorations were based on an 80 percent lung absorption of mercury vapor. Absorption of elemental mercury from the gastrointestinal tract probably less than 0.01 percent (87). Skin absorption is low (9).

Hursh and coworkers (39), in a study of five male volunteers, measured absorption of mercury vapor through the forearm skin. On the basis of their measurements, and exposure assumptions (air concentration of 50 µg/m³, and a skin area of 18,000 cm²), these investigators calculated a mean uptake of 10.4 µg/day mercury by this route during an 8-hour period. Use of the maximum measured absorption rate would nearly double this daily rate. These investigators calculated that an inhalation exposure from this same atmosphere would have resulted in 40 times as much mercury absorption based upon a respiration volume of 10 m³ for an 8-hour workday and 80 percent absorption.

Clarkson and colleagues (10) estimated that for the general U.S. population "the dominant exposure (for elemental mercury) is to mercury vapor from dental amalgams." These authors, using other researchers’ intraoral mercury data (56-58,61), estimated the amount of elemental mercury absorbed from dental amalgam restorations to range from about 2.9 to 17.5 1ug/day (2,900 to 17,500 ng/day). By comparison, estimates of absorbed elemental mercury from ambient air range from 32 to 96 ng/day.

Table 11 shows estimated mercury absorption from various sources presented in this document. The estimated amount due to inhalation of elemental mercury from dental amalgam restorations is much larger than that from ambient air. It is in the range of that estimated for both organic mercury and inorganic mercury (Hg⁺⁺) obtained from food (10,12). On the basis of data from Berglund and colleagues (33), it is also similar to the uptake of inorganic mercury (Hg⁺⁺) ingested in saliva.

Human Exposure and Response

Several epidemiologists have investigated mercury exposure and its resultant health effects. Most of these studies have focused on occupationally exposed workers who were not in the dental health field. Generally, results of dental health workers have shown that these workers have higher levels of mercury exposure than the general population, but lower levels than those of other occupationally exposed workers.

Occupations with High Mercury Exposure

Langworth and colleagues (93) measured the blood and urine mercury levels of 79 industrial workers who were not occupationally exposed and 89 chloralkali, workers who were occupationally exposed to elemental mercury. Levels of the workers' fish consumption was determined by questionnaire. The blood and urine mercury levels were 4 and 11 times higher, respectively, among occupationally exposed workers. They attributed the lesser differences in blood mercury levels to interferences from dietary exposure to methylmercury, which preferentially binds to red blood cells. Among unexposed workers, they found correlations between blood mercury levels and fish consumption and between urine levels and the number of amalgam surfaces. They did not find such correlations in subjects occupationally exposed to elemental mercury.

To determine whether short-term memory (STM) loss is a neurotoxic effect of exposure to elemental mercury, Smith and coworkers (88) conducted a study of mercury cell chloralkali workers from four plants. Two measures of STM were used, the Wechsler¹ digit span forward test, and a second test used to estimate a worker's 50 percent threshold for correct serial recall. The performance of 28 workers on the Wechsler digit span forward test did not show any association with urinary mercury exposure indices: the average urinary mercury concentrations for 3, 6, 12, and 24 months before testing. The average urine mercury values for these workers' exposure indices were 175, 164, 139, and 119µg/L, respectively.

In two separate plants, different from those in the previous study, 60 volunteers also took the 50 percent threshold span test. Reductions in STM capacity related to the 12-month average mercury-in-urine level (mean 102 µg/L) similar to those in the fist plant were observed among those with higher urinary mercury levels. The authors cautioned that the size of the reduction in memory span observed in these studies may be underestimated. The possible underestimation of effect results from some uncontrolled bias stemming from the inherently better STM of younger workers who because of their work assignment in the plants, also tended to have the higher urinary mercury levels. Although investigators used volunteers in this study and the populations studied were relatively small, the results were replicated in the two plants and are not likely to be the result of study bias.

In a cross-sectional study design, Levine and coworkers (92) evaluated peripheral nerve conduction tests on 18 workers at a chloralkali plant. Normal values were obtained from individuals aged 21 to 50 who were in good health with no known neurological deficit. Ulnar motor nerve normal values were obtained from 138 subjects and ulnar sensor nerve normal values from 82. The 18 subjects volunteering for the study were asymptomatic, and results of routine physicals conducted by the industrial physician at the time of the study were normal. Integrated mercury exposure was evaluated by averaging urine mercury concentration for the exposed subjects from the results of monthly urine tests from the previous 3 years. The mercury exposure indices covered from 3 to 36 months. Sensory distal latency correlated significantly with more than half of the urine mercury exposure indices used. Motor distal latency also showed significant correlation with mercury indices. These manifestations of toxicity were not apparent through standard physical examinations.

This plant's mercury control program removes workers from exposure when their spot urine mercury concentration exceeds 500 µg/L. These investigators concluded that the results of their study did not differ substantially whether using as a measure of body mercury concentration the number of months that urine mercury concentration exceeded either 500 µg/L or 250 µg/L and that their results offered no support for a threshold effect in the peripheral nervous system. Thus the degree of peripheral nerve involvement may relate to mercury as quantified by time-weighted urine mercury concentrations. Although the sample size in this study was small, the study was apparently well conducted and the findings correlate with measurable subclinical effects at urine mercury levels below the threshold for clinical effect of exposure to elemental mercury.

Piikivi and Tolonen (91) used a cross-sectional cohort design to study the cerebral effect of long-term (mean exposure time, 15.6 years) low-intensity exposure to mercury vapor in a group of 41 chloralkali plant workers and a group of age- and sex-matched wood processing plant workers with no occupational exposure to mercury. The time-weighted average concentration of mercury in blood was computed from health records that covered the period from 1969 until the study began in 1987. These investigators used the workers' long-term average mercury level in blood and the ratio of mercury levels in air to mercury levels in blood reported in the literature (94) to estimate the exposure level of the exposed group. On the basis of this relationship, they reported that the long-term exposure level of the chloralkali workers studied had been about 25µg/m³.

These investigators believe that long-term average mercury levels in blood do not provide accurate estimates of average long-term exposure because of the short physical half-life of inorganic mercury in blood. Thus, they chose to use blood values of mercury obtained at the time of the physical examination to estimate the long-term exposure of the persons studied. On the basis of such single blood samples, the long-term average exposure was estimated as about 25µg/m³ for day workers and 15µg/m³ for the shift workers. Comparison of computer-supported evaluation of EEGs obtained from mercury exposed and control workers showed those from the exposed group were significantly slower and more attenuated. This difference was most prominent in the occipital region, became milder parietally, and was almost absent frontally. Although the investigators found no concentration-response relationship, they concluded that the slowing and attenuation of computer-based evaluations of EEG observed in workers related to chronic exposure to elemental mercury at or below the current level that WHO recommended not be exceeded. It is likely, however, that the choice to define the long-term exposure on the results of a single blood sample at the time of study underestimated the actual long-term occupational exposure. This historical exposure was most likely at higher mercury vapor concentrations than those that currently exist in modem chloralkali plants. Hence the EEG effects they observed may have been the result of higher long-term mercury vapor exposures than the 25µg/m³ they estimated.

Because concentrations in neither blood nor urine accurately predict the concentration of mercury in the brain, there is little reason to believe that either measure reflects a dose-response relationship with EEG alterations. Results of this study, however, suggest that computer-based evaluations of EEGs may provide relatively early evidence of subclinical indices of neurotoxicity, but because the results also indicate that mental strain caused by shiftwork exacerbated disturbances in the EEGs, this method of evaluation should be used only in carefully controlled situations.

Rosenman and colleagues (89) performed a cross-sectional study using 42 workers employed at a plant producing inorganic mercury compounds. Mercury in blood and urine were sampled 4 months before and at the time of testing. A questionnaire survey was used to investigate worker's symptomatology, and standard neurobehavioral, saccadic eye movement, ophthalmologic, kidney function, and peripheral nerve conduction tests were performed and evaluated. Several subjects did complain of pain or numbness in their extremities. Although this does not negate the implication that peripheral nerve effects from mercury may occur, the experimental design was flawed.

The investigators found the following: 1) neuropsychological symptoms correlated with urinary mercury and urinary n-acetyl G -D -glucosaminidase (NAG) levels; 2) urinary NAG enzyme levels were increased compared with those of historical controls and were correlated with urinary mercury levels; 3) mean motor nerve conduction velocity was slowed and was associated with increases in urinary mercury levels and symptoms of numbness or pain in the extremities; 4) all mercury workers examined had lenticular opacities, but while there was no correlation with mercury level in blood or urine, these opacities are not found in the general population; 5) more abnormalities were seen in the saccadic eye movement of mercury workers than of the historical controls, but there was no relationship with mercury levels; 6) neurobehaviorial test scores were decreased compared with scores for the historical controls, but were not correlated with mercury levels.

Several factors make the results of the study by Rosenman and coworkers difficult to interpret. The questionnaire used in the study was not standardized and the results were correlated with mercury in urine not blood. More information on nerve conduction results is needed; furthermore, the symptoms are difficult to interpret. The eye findings for this small sample size show the following: 1) there was no relation to mercury levels, 2) the relationship to time on the job is difficult to interpret because the investigators found opacities in all workers examined, and 3) no specific control subjects were in the study. Findings on saccadic eye movement are not appropriately supported by references and do not involve specific control subjects. The subtle neurologic changes described are difficult to interpret because the information given to support the findings is not sufficient, or because the questionnaire used to ascertain symptoms was not sufficiently standardized.

Dental Occupations

Shapiro and colleagues (90) evaluated the potential for a relationship between cumulative exposure to mercury and chronic health impairment among 298 male dentists. The authors measured the mercury levels at the temple and wrist by x-ray fluorescence. The investigators reported neither exposure nor excretion data for these dentists. Peripheral nerve function was measured by standard electrodiagnostic methods. Neuropsychological tests used were the Wechsler Adult Intelligence Scale (WAIS) and the Bender-Gestalt finger-tapping and grooved-pegboard tests. Two independent judges blinded to the subjects' mercury levels, scored the Bender-Gestalt test results. For more than two thirds of the dentists, levels of mercury in the temporal region and the wrist were below the level detectable by the x-ray fluorescence technique (20µg/g of tissue). About 13 percent of the dentists had mercury values at the temple above 40µg/g.

Electrodiagnostic and neuropsychological findings for 23 dentists with mercury levels of more than 20µg/g tissue mercury levels were compared with those of a control group consisting of 22 age-matched dentists with no detectable mercury levels. Eight (30 percent) of the 23 dentists with more than 40µg/g Hg/g in the temple had polyneuropathies. The control group had no individuals with polyneuropathies. This finding was statistically significant (p = 0.008).

Electrodiagnosis is a sensitive technique that often detects subclinical neuropathies in apparently "normal" persons. The group with detectable mercury also had mild visuographic dysfunction and more self-reported distress symptoms than did the control group. Despite these dysfunctions all subjects in this study were apparently performing their professional tasks adequately and did not show intellectual impairment. Results of this well-designed and well-executed study showed that those dentists who used dental restorative materials containing mercury and had detectable tissue levels of mercury also had measurable biological dysfunctions.

Nilsson and coworkers (95) studied 505 persons occupationally exposed to mercury while working at 82 dental clinics in northern Sweden. Controls were 41 persons without occupational exposure to mercury. These groups were compared for urine mercury levels and symptoms previously associated with mercury exposure. The median mercury concentration in the workplace air was 1.5µg/m³ for workers in public dental care and 3.6µg/m³ for workers in private dental care. The investigators did not explain why the air values obtained during this study were lower than the range of 20-40 µg/m³ reported from other investigations of dental offices (96-100).

Urine mercury levels of dental personnel in the study by Nilsson and colleagues were similar to those for the general Swedish population. The investigators concluded that for those individuals occupationally exposed, the estimated burden of mercury from their own amalgam fillings was similar to the burden from the working environment. The prevalence of the four symptoms investigated (i.e., loss of appetite, tremor, insomnia, and anxiety) was less than 11 percent for both exposed and unexposed study subjects. Results of this study also showed no increase in the prevalence of these symptoms in relation to concentrations of mercury in urine.

Nilsson and coworkers concluded that the inconsistency between the findings of their study and of other studies (90), which did show an association between occupational exposures in dental offices and neurotoxic effects may be due to two factors: 1) the generally lower levels of occupational exposure to mercury in Swedish dental offices, and 2) the less sensitive measures of neurotoxicity that they (Nilsson and colleagues) used. In addition, these findings are difficult to evaluate because (Nilsson and colleagues) did not use standardized questionnaires.

Naleway and colleagues (101) reported findings from on-site screenings at the American Dental Association 1985 and 1986 annual sessions. These screenings were a part of the health screening program (HSP) to identify dentists having elevated concentrations of mercury in their urine. Data generated from this study were used to examine the relationship between elevated urinary mercury levels, occupational exposure and kidney dysfunction. Measurements of concentrations of beta 2 microglobulin in serum and urine and of creatinine in serum, and also of creatinine clearance were used to evaluate kidney dysfunction. The mean urinary mercury values found in the 1985 and 1986 HSP were 5.8 µg Hg/L and 7.6 µg Hg/L, respectively. For about 10 percent of the subjects in the 1985 and 1986 studies, urinary mercury concentrations were above 20 µg/L. No clear relationship was demonstrated between elevated urinary mercury concentrations and kidney dysfunction. The reported absence of a clear relationship between urinary mercury concentrations and potential kidney dysfunction is in agreement with other findings at the mercury concentrations measured.

Information on the professional exposure of the subjects in the 1985 and 1986 HSP was obtained by questionnaire. A follow-up questionnaire that addressed psychological and neuropsychological symptoms was provided to participants who had elevated urinary mercury concentrations in the 1985 HSP. Analysis of responses to these questionnaires provided three significant relationships none of which were health effects. These relationships were associated with the following: 1) mercury/amalgam handling and skin contact, 2) the number of amalgams placed by the dentist, and 3) the number of hours of practice per week

Hypersensitivity

If patients are sensitized to any of the components of amalgam dental restorations or to their corrosion/degradation products, they may have allergic reactions. Most concern about hypersensitivity has focused on the mercury component of amalgams. Mercury is documented as an allergen (102,103); however, other components of amalgams may also be involved.

A standardized dental test series has been developed to screen for contact allergy to dental materials (104). Using this screening technique, which includes 21 chemicals, 955 patients with a tentative diagnosis of contact allergy to dental materials were tested at 16 dermatological clinics. Results of this screening show that of this group less than 2 percent tested as allergic to elemental mercury (105). Only a small proportion of mercury-sensitized individuals respond adversely to the placement of amalgam restorations. The few case reports of adverse allergic reactions to amalgam involve skin reactions, such as rashes and eczematous lesions, which sometimes occur as tele-reactions, that is, reactions occurring a distance from the initiating site. Gingivitis and stomatitis may also occur.

Vernon and coworkers (106) have reviewed 41 published cases of allergy to amalgam. The reactions occurred 2 to 24 hours after the amalgam was placed. Some of these cases went into remission even though the patient was not treated, but most cleared up only after the amalgam was removed. Immediate hypersensitivity reactions (Type I) have also been reported after amalgam restorations have been removed and inserted (107,108). Documentation of the etiology of the effect is often missing, and verification is often based only on the remission of the lesions following removal of the amalgam.

Olstad and colleagues (63) studied the relationship of urine levels of mercury and non-specific allergies and school absences among 73 sixth-graders. They found no association between urine levels and either parameter. Both measures, however, were non-specific and based on self-reported information. Therefore, the findings are inconclusive with regard to allergic sensitivities.

Psychological Outcomes Associated with Mercury Levels in Body Fluids and Tissues

Siblerud (109) studied urinary mercury levels and the mental status of 50 college student volunteers with amalgams and 51 with no dental fillings. The reported mean level of mercury, 3.70 µg/L, for the amalgam group was 201 percent higher than the 1.23 µg/L mean level for the nonamalgam group. The mercury value reported for hair samples from the amalgam group was 1.43 µg/g, or 26.5 percent higher than the comparable 1.13 µg/g for the non-amalgam group. Both differences, however, were statistically significant. Among those with amalgams, the number of fillings correlated with both urinary mercury (r = 0.46, p = 0.001) and hair mercury (r = 0.23, p = 0.09), supporting results previously reported (74,91). Because hair reflects dietary intake of methylmercury, its analysis will not likely provide a very useful tool in evaluating exposure to mercury vapor from amalgam dental restorations.

The students were also given a health questionnaire to complete at home and another to complete while waiting for the laboratory testing. Responses to the first questionnaire showed amalgam subjects significantly less happy and having less "peace of mind" than the non-amalgam group. On the second questionnaire, the amalgam group reported more emotional symptoms and a lifestyle involving greater consumption of sweets, cigarettes, alcohol, and coffee, but none of the differences were statistically significant. No information was provided on the reliability and validity of the apparently non-standardized questionnaires used. Thus, reliable conclusions cannot be drawn.

In a supplementary survey, Siblerud sent one of his questionnaires to nearly 300 patients (average age, 40.4 years) whose amalgams had been removed. The 86 who responded were asked to list mental health symptoms for the year before the amalgams were removed and to evaluate the symptoms after their removal. Most of the patients were pleased with the results; they indicated improvement in emotional factors such as depression, irritability, and anger, and in general health status. However, without appropriate controls, no definitive conclusions can be drawn from this survey.

Findings of Mercury Residues in Neurological Patients with Alzheimer's and Parkinson 's Diseases

Mercury concentrations have been studied in many diseases, including Alzheimer’s, Parkinson's, Kawasaki's (110-113) and multiple sclerosis. The studies of Alzheimer's and Parkinson's disease patients are examples of studies of chronic diseases showing elevated concentrations of mercury in tissue and fluid samples compared with concentrations found in controls. Wenstrup and coworkers (112) compared trace element concentrations in the brains of 10 Alzheimer's disease patients with levels of these elements in the brains of 12 age-matched controls (ages 59-83). These investigators quantified 13 trace elements by neutron activation analysis and found 5 statistically significant differences relative to controls. Elements showing elevated concentrations in Alzheimer's patients were bromine in whole brain tissue and mercury in the microsomal fraction of cells. Those showing diminished concentrations were rubidium in whole brain, and in nuclear and microsomal cell fraction, selenium in the microsomal fraction of cells, and zinc in the nuclear fraction of cells. Tissue for this study was collected from the temporal cerebral hemisphere of patients and controls. The investigators concluded that the most important of the observed imbalances was the elevation of mercury among Alzheimer's disease patients.

Sources of mercury among these Alzheimer's patients are unknown. The investigators, however, suggested dental amalgams and environmental sources, such as seafoods as possible contributing sources, but they did not report data on the number of amalgams in the subjects studied. Several mechanisms were suggested by which mercury might alter brain function in Alzheimer's disease; for example, in rats, mercury has caused decreases in protein synthesis and in levels of RNA and DNA. These investigators (113,114) and others (115,116) have found a marked reduction of protein synthesis in the brains of Alzheimer's patients. They (113,114) suggest that an elevated concentration of mercury could inhibit protein synthesis, resulting in neuronal degeneration and cell death.

Investigators have reported differences of trace element concentrations in brain tissue of Alzheimer's patients (73,117,118) as compared to that of non-Alzheimer patients. At this time, no studies have demonstrated whether mercury deposition is a causal event or a result of the brain's degeneration from the disease itself (i.e., disease related changes in transport mechanism).

Ngim and Devathason (119) designed a hospital-based case-control study using 54 patients having Parkinson's disease and 95 controls matched for age. The results of the study show a positive correlation of the presence of mercury in blood, urine, and hair with Parkinson's disease. These investigators, however, did not present any information on either occupational exposure or the presence or absence of dental amalgams among the subjects, and thus did not control for confounding variables.

In one study, investigators found that persons with Alzheimer's disease have alterations in tissue levels of mercury and other trace elements, and another group found that Parkinson's patients had elevated levels of mercury in urine, blood, and hair. The results of neither study (112,119) can be interpreted as showing a causal relationship, and both require confirmation.

Conclusions

Mercury is a toxic substance. For high exposures, observed mostly in occupational settings, the severity of response correlates with the duration and intensity of the exposure. The relationship between the severity of response and the duration of exposure has, however, not been quantified at levels of exposure associated with dental amalgam restorations. In addition, subtle signs and symptoms of chronic mercury intoxication may not be found through routine physical examinations. The subtle changes previously described require special tests not commonly used in routine examinations—that is, nerve conduction studies, measurement of alterations in EEG, and measures of psychomotor functioning.

In studies in which investigators have measured the mercury concentration in intraoral and exhaled air among small populations of people with and without amalgams, estimates of human uptake of mercury vapor released from dental amalgams have ranged from 1.24 to 29 µg/day. Measurements of mercury in blood among subjects with and without amalgam restorations (61) and subjects before and after amalgams were removed (64,71) provide the best estimates of daily intake from amalgam dental restorations. These values are in the range of 1 to 5 µg. The blood mercury levels attributed to dental amalgams range from 0.4 1µg/L to 1.13 µg/L. Urine mercury levels are reported to be threefold to sixfold higher for those with amalgam fillings than for controls. Concentrations in tissue for those with amalgam fillings compared with those without amalgams are reported to be twofold to threefold higher in brain tissue and ninefold higher in kidney tissue.

Most data suggest that the daily mercury dose is 1 to 5 µg higher for subjects with 7 to 10 amalgams than for persons with no amalgams. Although specific data on subjects with recently placed fillings are scant, results of studies have shown that among these people levels of mercury in fluid spike after the fillings have been placed. No controlled clinical studies of health consequences have been conducted in association with the placement or removal of amalgam. Similarly, investigators have not looked for the subtle neurological and behavioral changes that have been demonstrated in some studies of occupationally exposed populations.

In low-level occupational exposures, the subclinical effects detected have occurred in groups with mean tissue mercury levels that are only tenfold higher than those of the general population; however, the relationship between the observed effects and the tissue levels is not clear.

Available data are not sufficient to indicate that health hazards can be identified in non-occupationally exposed persons. Health hazards, however, cannot be dismissed. Because there are no scientifically acceptable studies with sensitive, standardized measurements for physiological and behavioral changes in non-occupationally exposed populations, we cannot, at present, determine whether such changes observed in persons with low-level occupational exposure to mercury also occur as a result of exposure to mercury from dental amalgams.

The margin of safety may, however, be lower because of sensitivity to mercury or because body burdens of mercury are already high as a result of exposure to other sources; some persons may perhaps respond adversely to the incremental exposure to mercury derived from dental amalgams.

At the mercury doses produced by amalgam fillings, the evidence is not persuasive that the wide variety of non-specific symptoms attributed to fillings and "improvement" after their removal are attributable to mercury derived from the fillings. Conversely, the evidence is not persuasive that the potential for toxicity at the levels attributable to dental amalgams should be totally disregarded. The potential for effects at levels of exposure produced by dental amalgam restorations has not been adequately studied.

Research Recommendations

After review of the literature, the committee recommends that the following research be undertaken to clarify the effects of long-term, low-level mercury exposed from amalgam dental restorations.

Alternative materials should be tested for safety and effectiveness in animals and humans.

REFERENCES